Research On Virtual Three Dimensional Reconstruction System

In modern three dimensional visualization medical researches, the method of digital image processing combined with computer technologies applied in the field of medical imaging was an important research content. Slice is a powerful tool for detection and diagnosis of the liver, but even with lots of slice images, not only the inner blood vessels of the liver but also the relative position of tumor cannot display visually. As a result, diagnostic and surgical procedures become so time-consuming, easily missed or misdiagnosed, however,using visualization system is an effective and important way to improve the diagnosis efficiency and reduce some u unacceptable issues such as clinical misdiagnosis.Traditional virtual three-dimensional reconstruction system mainly focus on CT images with clearly boundaries, while using complex fully manual segmentation to these sequence diagrams with more subtle and unapparent boundaries, then three-dimensional reconstruction.In order to show the internal structure more accurate, it often requires image processing for amount of sequence diagrams, which lead to its low efficiency. So this paper focuses on such issues,conducted a series of researches about valid information extraction of liver and its inner complex blood vessels, 3D reconstructions, three-dimensional cutting and interactive measurement. In the end,we achieved a three-dimensional virtual reconstruction system in a more comprehensive level.First, according to the nature characteristics of medical image, we use the segmentation algorithm that combining seed growing region method and linear interpolation method to automatic segment the medical sequence images in batches innovatively. Meanwhile we read seed point and the pixel value of target area in real-time through adding mouse click callback event. And finally, using multi-thread method to improve the speed of image segmentation.Secondly, using marching cubes surface rendering algorithm to achieve three-dimensional reconstruction for segmentation sequence images, then display realistic and intuitive external shape and internal blood vessels of the liver on-screen. And proposed three-dimensional spatial filtering between adjacent layers method to remove excess noise points;Third, given the disadvantage of marching cubes algorithm and research several accelerate three-dimension reconstruction methods with comparative analysis for experiment results, and Roaming 3D scene with virtual camera is available.Fourth, cutting operation towards the results of three-dimension reconstruction with arbitrary or several cutting plane make the internal information available to operators.Meanwhile,using the combination method of bounding box and binary tree algorithm to optimize the system so that the real-time property improved.Finally, using pick-up method to achieve three-dimensional measurement, such as angle measurement and distance measurement as well as Irregular area and volume measurement.And researches about the measurement error were also carried out.Experiment results confirms that the virtual-dimensional reconstruction system proposed in the paper shows the more realistic outer contour and internal blood vessels of the liver model, as well as to achieve arbitrary cutting and three-dimensional roaming operations.Finally, carried out a series of researches about irregular area and volume measurement about the liver.